Virology 535 (2019) 59–73 Contents lists available at ScienceDirect Virology journal homepage: www.elsevier.com/locate/virology Genomic comparison of 60 completely sequenced bacteriophages that infect T Erwinia and/or Pantoea bacteria ∗ Daniel W. Thompsona, Sherwood R. Casjensb,c, Ruchira Sharmaa, Julianne H. Grosea, a Department of Microbiology and Molecular Biology, Brigham Young University, Utah, USA b Division of Microbiology and Immunology, Department of Pathology, University of Utah School of Medicine, University of Utah, Salt Lake City, UT, 84112, USA c School of Biological Sciences, University of Utah, Salt Lake City, UT, 84112, USA ARTICLE INFO ABSTRACT Keywords: Erwinia and Pantoea are closely related bacterial plant pathogens in the Gram negative Enterobacteriales order. Bacteriophages Sixty tailed bacteriophages capable of infecting these pathogens have been completely sequenced by in- Erwinia vestigators around the world and are in the current databases, 30 of which were sequenced by our lab. These 60 Pantoea were compared to 991 other Enterobacteriales bacteriophage genomes and found to be, on average, just over Bacteriophage twice the overall average length. These Erwinia and Pantoea phages comprise 20 clusters based on nucleotide and Phage protein sequences. Five clusters contain only phages that infect the Erwinia and Pantoea genera, the other 15 Cluster Diversity clusters are closely related to bacteriophages that infect other Enterobacteriales; however, within these clusters Erwiniaceae the Erwinia and Pantoea phages tend to be distinct, suggesting ecological niche may play a diversification role. Enterobacteriales The failure of many of their encoded proteins to have predicted functions highlights the need for further study of these phages. 1. Introduction possible bio-terrorism agents. For example, Pantoea agglomerans is the causative agent of potato blight and has also been documented as an Bacteriophages are viruses that infect bacteria. Their virions are opportunistic human pathogen (Sengupta et al., 2016), and Erwinia comprised of a protein shell containing genetic material that can be amylovora is the causative agent of fruit tree fire blight, which is re- dsDNA, ssDNA, dsRNA or ssRNA. Their genomes can contain as few as sponsible for an average of 100 million US dollars damage annually to 3.3 kb or as many as 500 kb (Fiers et al., 1976; Hatfull and Hendrix, apple orchards in the United States (Norelli et al., 2003). Fire blight 2011). They are the most abundant and diverse biological entities, with infections are currently treated with antibiotics; however, up to 70% of an estimate of about 1032 tailed bacteriophages on Earth (Suttle, 2007). these bacteria found in nature are resistant to the currently used anti- Since bacteriophages are parasites of bacteria, they have played an biotics (Forster et al., 2015). Due to their ability to kill their bacterial important role in the evolution of bacteria. Bacteriophages can have hosts, phages are projected to provide an alternative anti-bacterial two alternate lifestyles when infecting a bacterium, lytic and temperate. therapy for these plant diseases. Lytic phages simply replicate to form progeny virions which are re- A number of bacteriophages that infect Erwinia or Pantoea leased to infect other host cells. Temperate phages can also propagate (Erwiniaceae phages) have been isolated by a variety of investigators lytically but may instead enter a semi-dormant “prophage” state in from several continents, and 60 of their complete genome sequences are which the phage DNA either replicates as a plasmid or integrates into available at the National Center for Biological Information (NCBI) the host chromosome and replicates passively as part of that replicon. GenBank database (Sayers et al., 2019). Thirty of these phages were Prophages can be stable indefinitely, but environmental triggers can isolated and characterized in our laboratory (Esplin et al., 2017; cause their “induction” into the lytic growth cycle. Sharma et al., 2018). Host range studies of phages that infect Erwinia- Erwinia and Pantoea are very closely related Gram negative bacteria ceae suggest relatedness between the two host genera in that several in the Erwiniaceae family of the Enterobacteriales order (F.S. Dworkin phages isolated on Erwinia can infect both Erwinia and Pantoea strains et al., 2006) that are often plant pathogens, causing necrosis in the including phages Joad and RisingSun (Arens et al., 2018), Y3 (Buttimer tissues of the infected plant. These pathogens are a large burden on the et al., 2018), ØEa2809 (Lagonenko et al., 2015) and CBB (Buttimer agricultural community of the United States and are currently listed as et al., 2018). Herein, we compare these 60 bacteriophages and place ∗ Corresponding author. E-mail address: [email protected] (J.H. Grose). https://doi.org/10.1016/j.virol.2019.06.005 Received 16 April 2019; Received in revised form 5 June 2019; Accepted 11 June 2019 Available online 18 June 2019 0042-6822/ © 2019 Elsevier Inc. All rights reserved. D.W. Thompson, et al. Virology 535 (2019) 59–73 them in 20 different clusters based on their genomic and proteomic isolation location, many were found in infected trees or in the soil traits. An analysis of each of these clusters is provided, along with around them. Of the 30 phages we isolated, the genomes of only two, comparisons to known phages. The purpose of our analysis is four fold: Joad and RisingSun, have been fully discussed in the literature (Arens 1) to gain insight into the relationship and interaction between different et al., 2018), 26 have been reported in only genome announcements, bacteriophage types and their host bacterial species, 2) to further un- and two (Rebecca and Derbicus) are first reported here. derstand the relationships among members of the Enterobacteriales The genomes of the Erwiniaceae phages range from 378,379 bp order by comparing their bacteriophages, 3) to contribute to our un- (phage CBB) to 29,564 bp (phage ENT90). The average genome length derstanding of overall bacteriophage diversity, and 4) to provide in- of the 991 other Enterobacteriales tailed phage genomes currently in the formation that will aid in the treatment of the above plant diseases by NCBI database is 81,187 bp, but the Erwiniaceae phages have an average development of improved phage therapy cocktail design and safety. genome length of 162,734 bp. Thus, Erwiniaceae bacteriophages com- prise about five percent of the sequenced Enterobacteriales tailed 2. Materials and methods phages, and the average genome size is almost double the overall average. Fig. 1 plots the length of all the Enterobacteriales tailed phage 2.1. Isolation, sequencing and assembly of phages genomes and indicates the locations of the Erwiniaceae phages. The Erwiniaceae phage genome lengths are within the previously known Thirty Erwinia phages were isolated by our laboratory at Brigham extremes, but it is not known if their large average size is the result of Young and 28 of these have been previously described (Esplin et al., isolation methods used, properties of the hosts or the skew in isolation 2017; Sharma et al., 2018). The genomes of the two previously un- sources toward trees and the soil around them. described phages Rebecca (accession No. MK514281) and Derbicus The 60 Erwiniaceae phage whole genome nucleotide sequences were (MK514282) were sequenced from libraries made with the Illumina compared with Genome Pair Rapid Dotter (Gepard) (Krumsiek et al., TruSeq DNA Nano kit and Illumina HiSeq 2500 sequencing (250-bp 2007)(Fig. 2A). By the criterion of diagonal line strength, these phages paired end). Genomes were assembled with Geneious (Kearse et al., fall into 20 clusters that have similarity over 50% of the phage genome 2012) version 8.1 using de novo assembly with medium-low sensitivity as previously described (Grose and Casjens, 2014; Hatfull et al., 2010). as described previously in Sharma et al. (Sharma et al., 2018). Coverage The clusters in Fig. 2A are indicated by the founding Erwiniaceae phage depths were 527–1615 (1015.6 average) for Rebecca and 183–1758 in the group (the first sequence released in GenBank) unless the phage (572.3 average) for Derbicus, both phages circularized their genomes belongs to a previously-described Enterobacteriales cluster, in which upon assembly. case the previously published name for that cluster is used (Grose and Casjens, 2014). An Average Nucleotide Identity (Esplin et al., 2017) 2.2. Genomic analysis and comparison matrix was also constructed using Geneious (Kearse et al., 2012), and if phage clusters are defined so that each phage has ≥50% ANI withat Gepard (Krumsiek et al., 2007) was used to generate dot plots that least one other phage in the group and ≤24% ANI with phages from compare nucleotide sequences of multiple genomes. Default settings other clusters (Supplementary Table S1), the ANI grouping matches the (word size 10) were used to generate dot plots, however lower and dot plot-defined clusters perfectly. Our clusters correspond in general to upper color limit were increased in order to allow better image viewing. genera or subfamilies that have been defined by the International Geneious (Kearse et al., 2012) was used to align the sequences in an Committee on Virus Taxonomy (ICTV), but a number of our clusters identity matrix using MAFFT plugin and setting parameters to auto- have not yet been formalized by that group. algorithm, a scoring matrix of 200PAM/k = 2, a gap open penalty of In addition to genome nucleotide sequence analysis, whole pro- 1.53 and an offset value of 0.123. Phamdb (Lamine et al., 2016), an teome and single protein analyses support these 20 clusters. Whole online version of Phamerator (Cresawn et al., 2011), a bioinformatic proteome analysis was performed using Phamerator (Cresawn et al., tool designed to compare bacteriophage genomes was used to visualize 2011) to group the phage-encoded proteins into related “Phamilies”, both nucleotide and protein similarity using kClust (Hauser et al., and SPLITSTree (Huson and Bryant, 2006) was used to infer relation- 2013).